Zalos has secured $3.6 million to automate financial workflows.

      The startup Zalos, part of the YC Fall 2025 batch and co-founded by a former Agicap General Manager and a former Apple Pay engineer, transforms screen recordings of financial workflows directly into computer agents without the need for API integration. The funding round was led by 14 Peaks, with contributions from Cohen Circle and 20VC.

      The software stack used by CFOs both presents challenges and imposes limitations. Typically, enterprise finance teams utilize a mix of ERPs, CRMs, spreadsheets, emails, and banking platforms that were developed at various times, by different companies, for distinct purposes. The APIs connecting these systems are frequently incomplete or missing altogether, necessitating that finance teams bridge the integration gap manually by downloading, reformatting, uploading, and reconciling data across systems to perform tasks that should be automated.

      Zalos, which is based in San Francisco and London and emerged from Y Combinator’s Fall 2025 cohort, has secured $3.6 million with the belief that the solution lies not in a new ERP but in a novel type of agent that manages the existing stack as a human analyst would. The investment round was led by the Swiss venture capital firm 14 Peaks, with participation from Cohen Circle and 20VC. The angel list is particularly notable for its industry relevance, featuring individuals such as Mike Lenz, CFO of FedEx; Ian Sutherland, CFO of the UK business bank Tide; and Paul Forster, founder of Indeed, along with others experienced in finance software, accounts payable, and enterprise infrastructure.

      Zalos takes a notably straightforward technical approach. Instead of requiring API integrations or custom connectors, the company trains agents using screen recordings of the actual workflows that finance teams perform using their current tools. For example, a billing cycle in NetSuite, a reconciliation in SAP S/4HANA, or a month-end close in Sage serves as the training input. The agent then mimics that process, logging in with credentials, navigating through screens, entering information, and managing two-factor authentication, all without altering the underlying system. Every action is documented in an auditable log, and the platform has achieved SOC 2 Part II certification. The key commercial insight is that many enterprise automation efforts in finance falter due to absent APIs, insufficient data exposure, or prolonged integration efforts before any functionality is realized.

      The two founders reached a similar conclusion from different paths. CEO William Fairbairn previously served as UK General Manager at Agicap, a CFO-targeted software firm valued around $800 million, where he engaged in numerous discussions with finance leaders who expressed common frustrations regarding ERP implementations: lengthy projects with modest benefits when successful and significant career risks if they fail. CTO Hung Hoang worked at Apple for five years, developing Apple Pay’s Buy Now Pay Later product and other AI initiatives, and became interested in computer agents partly through his work at Twin, a lab focused on this technology. The two met at Y Combinator and started building Zalos in October 2025.

      Zalos has a clear yet competitive market positioning. OpenAI's Operator and Anthropic's computer use functionalities work at a general-purpose level and can perform tasks across various interfaces. In contrast, Zalos bets that finance operations necessitate high levels of accuracy, audit trails, and domain-specific skills (like Excel manipulation, ERP navigation, and categorization logic) that general-purpose agents cannot consistently deliver. The company currently serves midmarket and enterprise finance teams, with plans to broaden its reach into additional enterprise ERPs and on-premise systems with the new funds.